Background: Several attempts to classify dyspepsia into subgroups have been proposed as a basis for empirical treatment and research. However, subgrouping has proved difficult due to overlap of symptoms between subgroups, and the response to empirical therapy is difficult to predict. We aimed to study whether natural symptom combinations occur in patients seeing general practitioners because of dyspepsia and whether symptom presentation could predict the effect of proton pump inhibitor treatment.
Methods: The symptom presentation of 7270 consecutive, unselected patients with dyspepsia in general practice was studied by using principal-components analysis. The relation to the effect of omeprazole was studied in a subsample (n=471) with predominantly reflux-like or ulcer-like dyspepsia being included in a controlled clinical trial of omeprazole versus placebo.
Results: Four principal components (factors), explaining 36% of the total variance, were found. They describe four independent dimensions in the symptoms of dyspepsia that can be interpreted meaningfully as representing A) acid-related disease of the upper gastrointestinal tract, B) irritable bowel disorder, C) dysmotility of the stomach/duodenum, and D) dysmotility of the esophagus. In the subsample the response to proton pump inhibition therapy was associated with high component-A scores, low component-B scores, and low component-C scores. A pocket chart was devised to obtain the component scores easily in new patients.
Conclusion: The analysis identified four characteristic, biologically meaningful dyspepsia components that express independent dimensions in the symptoms of patients with dyspepsia. The symptom scores corresponding to the four components may improve symptom-based diagnosis and thereby empirical therapy. In particular, the association between component scores and the effect of omeprazole suggests that classifying dyspepsia on the basis of these components may focus empirical omeprazole therapy even more.